首页|期刊导航|大气和海洋科学快报（英文版）|A deep learning-based U-Net model for ENSO-related precipitation responses to sea surface temperature anomalies over the tropical Pacific
大气和海洋科学快报（英文版）2023，Vol.16Issue(4)：57-64,8.DOI:10.1016/j.aosl.2023.100351
A deep learning-based U-Net model for ENSO-related precipitation responses to sea surface temperature anomalies over the tropical Pacific
A deep learning-based U-Net model for ENSO-related precipitation responses to sea surface temperature anomalies over the tropical Pacific
摘要
Abstract
SST-降水反馈过程在热带太平洋ENSO演变过程中起着重要作用,能否真实地在数值模式中表征SST-降水年际异常之间的关系及相关反馈过程,对于准确模拟和预测ENSO至关重要.例如,在一些模拟ENSO的混合型耦合模式中,通常采用大气统计模型(如经验正交函数;EOF)来表征降水(海气界面淡水通量的一个重要分量)对SST年际异常的线性响应.然而在当前的耦合模式中,真实观测到的降水-SST统计关系还不能被很好地再现出来,从而引起ENSO模拟误差和不确定性.在本研究中,使用基于深度学习的U-Net模型来构建热带太平洋降水异常场对SST年际异常的非线性响应模型.研究发现:U-Net模型的性能优于传统的基于EOF方法的模型.特别是在热带西太平洋海区,U-Net模型估算的降水误差远小于EOF模型的模拟.此外,当SST和降水异常的趋势信息作为输入变量也被同时引入以进一步约束模式训练时,U-Net模型的性能可以进一步提高,如能使热带辐合带区域的误差显著降低.关键词
U-Net模型/EOF方法/SST-降水年际异常关系/CMIP6模拟Key words
U-net model/EOF method/SST-precipitation feedback/CMIP6 simulations引用本文复制引用
Yuchao Zhu,Rong-Hua Zhang..A deep learning-based U-Net model for ENSO-related precipitation responses to sea surface temperature anomalies over the tropical Pacific[J].大气和海洋科学快报（英文版）,2023,16(4):57-64,8.基金项目
The first author was supported by the National Natural Science Foun-dation of China[grant number 42276008]and the Laoshan Laboratory[grant number LSKJ202202403-2].The second author was also sup-ported by the National Natural Science Foundation of China[grant num-ber 42030410],as well as the Strategic Priority Research Program of the Chinese Academy of Sciences[grant number XDB40000000]and the Startup Foundation for Introducing Talent of NUIST. （）
